Maine State Library Maine State Documents Alcoholic Beverages and Lottery Operations Administrative and Financial Services Documents 9-2014 Maine Lottery Segmentation Study, 6 Segment Solutions – Segment Overviews Simon Jaworski Leger Lance Henik Leger Follow this and additional works at: http://digitalmaine.com/bablo_docs Recommended Citation Jaworski, Simon and Henik, Lance, "Maine Lottery Segmentation Study, 6 Segment Solutions – Segment Overviews" (2014). Alcoholic Beverages and Lottery Operations Documents. Paper 4. http://digitalmaine.com/bablo_docs/4 This Text is brought to you for free and open access by the Administrative and Financial Services at Maine State Documents. It has been accepted for inclusion in Alcoholic Beverages and Lottery Operations Documents by an authorized administrator of Maine State Documents. For more information, please contact [email protected]. Maine Lottery Segmentation Study 6 Segment Solutions – Segment Overviews Prepared by: Prepared for: Simon Jaworski Sr. Vice President, Leger Maine Lottery Lance Henik Scientific Games Sr. Account Manager Background and Objectives As with any market research endeavors, actionability is the key to any segmentation approach! To be successfully actionable, a segmentation study must address the following three key criteria: Who They Are? What Lottery Players Do? Why They Do It? To this end, Leger’s objectives for the Maine Lottery Segmentation Study were to: 1. Allow the Lottery to use information attained in the segmentation study for strategic planning, product development, product enhancements, pricing, market positioning and communications. It is anticipated that this data will help the Lottery maximize revenue potential and minimize risk associated with new product introductions and product enhancements. 2. Learn the values, attitudes, motivations and lottery buying behavior of Maine adults which results in a division of the public into categories known as segments. 3. Find segments that could be easily understood yet differentiated from each other. Factors assessed but were not limited to: geographic, demographic, 2 socioeconomic, lifestyles, motivations and product usage. Leverage This Segmentation To Its Fullest Potential! The Leger team believes the Maine Lottery has a tremendous opportunity to maximize the investment made and to leverage the insights garnered from this study. There are two primary considerations with regard to leveraging this segmentation study: o Buy-in The key to a successful segmentation is internal ‘buy-in’ from as many departments as possible within an organization, on both the front-end and back-end of the segmentation as possible. Leger’s most recent presentation with the Maine Lottery, which took place in June 2014, ensured this took place with the Lottery Director, along with heads of Marketing, Sales and Research teams, all having attended the initial segmentation session. The Maine Lottery needs to embrace the results from top to bottom in the organization. o Maximizing Value and Opportunity The segmentation should also maximize the opportunity for talking to the current (and future) base of Maine Lottery players. This research offers the chance to talk to a wide ranging audience about potential new products, advertising and communications that are already in the pipeline. This opportunity should not be passed up. 3 Methodology Maine respondents participated in a 40 minute survey between March 5 and March 27, 2014. 1,013 completed the online survey The respondents all had to: Be older than 18 years of age Have no moral objections to Maine running a state lottery Primary (or secondary) resident of Maine Not work for an advertising agency, market research company, marketing consultant, any type of Lottery or a store that sells lottery tickets Margin of Error – Statistical Sample Accuracy at 90% confidence on n=1,013 completes is ± 2.6% 4 Project Review Leger has based the timeline on a decision being made the week of August 18th 2014 Data *Review of Approval of Final Maine Lottery Field period Project Questionnaire analysis, Segment Options design and Approval, Pre- March 5th - building of via meeting at Test, and Questionnaire March 27th the Segments Maine Lottery 2013 Study Launch Programming April June 3rd Provide the relevant short- Final Initial Finalize Decision Segmentation form version(s) Segmentation Segmentation Names of the new Made on report Naming and Delivery of Final Segmentatio Maine Lottery n Solution Delivered August 2014 Segmentation Report segments th July 7th July 15 September 2014 September 18th 5 Segmentation Specifics (1) o Leger utilizes exhaustive search criteria to surface segments which are more projectable: . Use broad . Add additional variable classes class and evaluate (attitudinal, improvement, fit . Identify variables . Determine model behavioral, etc.) and parsimony for modeling fit and parsimony in models to examine initial clustering results. Model Run Model Run Model Run Cont’d . Evaluate cluster . Develop Re-iterate for profiling and better fit composite model statistics variables as well as alternative Discriminating numbers of Variables clusters Discriminating . “Bootstrap L2” – . Compare cluster Classes random iterations membership and to re-estimate the . Pull out non- p-value profiling across informative the exploratory variables results (e.g., how Factor/Principle . Component Analysis well Apply known behaviorally- class based clusters indicators align with attitudinally- based clusters). 6 Segmentation Specifics (2) Leger aimed to surface segments that are: o Commercially significant, by being: o Large enough o Differentially responsive to potential marketing/sales programs and new product launches (such as new Instant Games) o Strategically actionable o Distinct and reproducible – avoiding artifacts of the analytic process o Identifiable and accessible – tactically actionable . Size and characterize the . Evaluate the statistically valid segments refined model . Discrimination Model Run Cont’d . Use business . We use multiple techniques to knowledge to surface key differentiating assign characteristics among segments unclassified cases including: . ANOVA + . Discriminant analysis - . CHAID 7 Segmentation Technical Overview (3) o Considering the research objectives, Leger had the flexibility to recommending several segmentation options. o The decision of exercising latent class segmentation or traditional clustering was made after data collection, so that the team could maximize the methodology to surface meaningful segments. o Latent class clustering, Convergent Cluster Ensemble Analysis (CCEA), and traditional methods such as K-means clustering were used during the course of analytics. o The final solution, however, was based on the CCEA segmentation – which involves a two step process. 1. Develop multiple cluster analyses that vary in terms of the clustering method employed (k-means, Hierarchical clustering using the number of clusters (ranging from 2 to 30) and the measures used. 2. Create ‘clusters’ and then ‘group’ respondents based on the analyses generated in the first step. While different Meta-Clustering Algorithms can be used to cluster respondents, CCEA Convergent Cluster & Ensemble Analysis uses k–means (a distance-based algorithm) to cluster respondents and create a consensus solution. A. This involves iterating from random but strategically chosen starting points; CCEA replicates each analysis ~30 times. B. Each replication is compared to every other to assess its reproducibility, and the most reproducible solution is selected. o The advantages of CCEA include: 1. Allows to combine groupings from alternate and dissimilar sets of variables (e.g., demographics, lifestyle batteries, desired benefits or needs, etc.); 2. Includes a variety of clustering techniques when building the ensemble in the first step (1 ); 3. Incorporates legacy clusters that are based on internal data; 4. Uncovers better, more robust cluster solutions that are less sensitive to sample variations and outliers; and 5. Allows to find solutions that would not have been uncovered using a single approach. 8 References (for additional information): Retzer, J. and M. Shan (2007), “Cluster Ensemble Analysis and Graphical Depiction of Cluster Partitions,” Proceedings of the 2007 Sawtooth Software Conference, Sequim WA. Sawtooth Software (1998), “CCA System,” Evanston, IL. Strehl, A. and J. Ghosh (2002), “Cluster Ensembles — A Knowledge Reuse Framework for Combining Multiple Partitions,” Journal on Machine Learning Research (JMLR), 3:583-617, December 2002. Segmentation Specifics (4) o A Value-based segmentation (“VBS”) can help facilitate the Maine Lottery to create a five stage output, such as the one identified below, that will add value at every level of its business model: Prioritize marketing efforts according to segment potential on Identify segment(s) that demands “what works with them” in terms top-line. The Maine acquisition, retention, or defection of marketing messaging and Lottery can now prioritize strategy. segments, based upon strategic This is particularly important when positioning managing goals – expanding player base, The Maine Lottery now has the and addressing competitive forces. opportunity to sculpt Marketing and increasing purchase frequency, etc. PRIORITIZATION Messaging plans by Segment. ST G R IN AT T AC E Project segment KE T GI Actionable and R IO C MA N membership of targetable customers and non- segments. Leger customers – Critical selected the optimal